INTELLIGENCE ARTIFICIELLE - Peut-elle remplacer le philosophe ? — Transcript
Full transcript
- 0:00Before leaving you with your episode, I
- 0:03would like to remind you, or let you
- 0:05know if you didn't know, that my book
- 0:07Philosophy is for You Too is available
- 0:09for pre-order on Amazon and on the Fnac
- 0:12website.
- 0:14A book published by Larousse, 240 pages
- 0:17, € 19.99.
- 0:20A book in my style: clear, concrete,
- 0:23accessible.
- 0:25If you appreciate my work, you can't
- 0:27miss this book.
- 0:29Pre-order it today. The links are in the
- 0:32description. On that note, let's get to
- 0:34your episode.
- 0:38Hello everyone.
- 0:41For the past few years, artificial
- 0:44intelligence has been making its way
- 0:46into all areas: industry, health,
- 0:48artistic creation, and even philosophy.
- 0:53We ask it to comment on Plato, summarize
- 0:56maps, explain Hegel, or compare
- 0:59Nietzsche, and it does so in a matter of
- 1:02seconds with a fluidity and clarity that
- 1:06can make your head spin.
- 1:10So, the question arises: Can artificial
- 1:12intelligence eventually replace
- 1:14philosophers?
- 1:18Aristotle said that philosophy begins
- 1:21with wonder.
- 1:23And this wonder is born of experience,
- 1:27an experience of the world, of death, of
- 1:30injustice.
- 1:33It is because man suffers that he begins
- 1:35to think.
- 1:37A machine does not suffer.
- 1:41It does not know what it means to lose a
- 1:44loved one, to desire without being
- 1:46fulfilled, or to face the absurd.
- 1:51Thinking is not just processing
- 1:54information, it is not piling up data or
- 1:57recombining sentences.
- 2:01Thinking is questioning, it is
- 2:04confronting the unknown, the uncertain,
- 2:07the uncomfortable.
- 2:11As Simone Veille wrote, thinking is
- 2:14saying no.
- 2:17An artificial intelligence, despite its
- 2:20name, is not intelligent.
- 2:23It calculates, it predicts, it imitates
- 2:26, but it experiences nothing.
- 2:31When it explains Kant's metaphysics to
- 2:33you, it does not understand what it is
- 2:35saying, it reproduces it.
- 2:38But let's go further.
- 2:40Let's imagine that one day an animal
- 2:43becomes conscious, that it has emotions
- 2:46, that it no longer just imitates.
- 2:51Perhaps one day it will succeed and
- 2:53perhaps then we will be surpassed?
- 2:57But until we are, there is one thing
- 2:59left for us to do.
- 3:02We must not think about delegating this
- 3:06task to algorithms, but about embodying
- 3:10it, living it and perhaps, who knows,
- 3:13making it something human.
- 3:17Thank you.
- 3:19Hop hop hop, wait a minute, not so fast
- 3:22.
- 3:23What just happened?
- 3:26Well, what just happened was that what
- 3:29you just listened to wasn't me, it was
- 3:32Lia.
- 3:35The text was Lia's and the voice was
- 3:38also Lia's.
- 3:42Pretty good, huh?
- 3:44I have just one question for you.
- 3:47Did you believe it?
- 3:50I mean, did you believe it?
- 3:53Not now that I've mentioned it.
- 3:55Are you thinking, "Oh yeah, there was
- 3:58something weird going on."
- 4:01No, what I want to know is, did you
- 4:03believe it at the time?
- 4:07Tell me in the comments because if you
- 4:10believed it, well that means that to the
- 4:13question "Can Lia replace the
- 4:15philosopher?"
- 4:17There's a good chance that the answer is
- 4:19yes.
- 4:22So, what I propose is that we take a
- 4:25look back at this experiment, that I
- 4:27explain how I proceeded and that we
- 4:29think a little about the questions it
- 4:32raises.
- 4:34Now, I'll make it clear right away that
- 4:36I'm not an AI specialist at all.
- 4:39I have no particular expertise on the
- 4:41subject, even though it's obviously a
- 4:43subject that interests me.
- 4:45But if you want to hear from someone who
- 4:48really masters the subject, I recommend
- 4:50Mr. Fi's videos.
- 4:52I think that if there's a popularizer of
- 4:55philosophy who has real expertise in
- 4:57artificial intelligence, it's him.
- 5:00In fact, he's one of the only ones to
- 5:02distinguish between AI and what are
- 5:04called LLMs.
- 5:06LLMs are language models like chat GPT
- 5:10or cloud.
- 5:12They're also called conversational
- 5:14agents.
- 5:16that is to say, text generators.
- 5:19So all this to say that I am not going
- 5:22to launch into a presentation on Lia's
- 5:24capabilities or limitations.
- 5:27What I am going to do instead is to
- 5:29share with you my personal thoughts on
- 5:31the question of whether Lia can replace
- 5:33the philosopher?
- 5:35A thought that the 3 minutes of fake
- 5:37Charles that you have just heard will
- 5:39provide a starting point for.
- 5:42So we are going to divide things into
- 5:45two parts.
- 5:46The first part is the technical part.
- 5:49How did I actually generate these 3
- 5:52minutes of fake Charles?
- 5:55We will try to go through this technical
- 5:57part fairly quickly, even if I really
- 5:59need to take the time to explain it to
- 6:01you.
- 6:02And then, we will move on to a second
- 6:04part, the one that I think interests you
- 6:07the most, namely the philosophical part
- 6:10.
- 6:10What can we say about this experience on
- 6:13a philosophical level?
- 6:15So, on a technical level, we will start
- 6:18with the voice.
- 6:20So for the voice, I used an application
- 6:23called Eleven Labs.
- 6:26Eleven Labs is a voice library where you
- 6:29will find two types of voices: cloned
- 6:31voices and synthesized voices.
- 6:35Synthesized voices are voices entirely
- 6:38generated by Ria, so entirely artificial
- 6:41if you prefer.
- 6:43While cloned voices are voices generated
- 6:46from a sample of a human voice.
- 6:50So, in concrete terms, these cloned
- 6:52voices are people who have recorded
- 6:54their voices and imported the audio file
- 6:56of their voice into the application.
- 6:59Once the file is imported, the
- 7:02application scans the voice and from
- 7:04there, it will be able to reproduce this
- 7:07voice and make it say whatever you want
- 7:10it to say.
- 7:12So, what I did is that I imported my
- 7:14voice into Eleven Labs and that wasn't
- 7:17difficult since I only had to import
- 7:19about ten of my episodes in audio.
- 7:23Levenlabs ended up with several hours of
- 7:25samples of my voice, which they scanned
- 7:28and I ended up with a voice that which I
- 7:30was able to make say whatever I wanted.
- 7:35So quite honestly, and I don't know what
- 7:37you think, but personally I find the
- 7:39result to be truly impressive.
- 7:41I find it really hard to tell that it's
- 7:44not really me.
- 7:46But I have to say that if the result is
- 7:49so impressive, it's mainly because I had
- 7:52to start over a lot.
- 7:54What you heard isn't the first result I
- 7:57got.
- 7:59I regenerated the sentences several
- 8:01times until I found the right intonation
- 8:03each time.
- 8:05Because what happens when you ask Eleven
- 8:07Labs to generate a sentence is that the
- 8:10intonation of the voice is placed
- 8:12somewhat randomly.
- 8:15Random in the sense that Lia isn't yet
- 8:17at the stage where she understands what
- 8:19she's saying, which means she doesn't
- 8:21put the voice inflections in the right
- 8:23places.
- 8:25I'll let you listen to it so it'll be
- 8:27clearer.
- 8:28There you go. So Here we are on Eleven
- 8:30Labs. I'll pick a random sentence. For
- 8:33example, can artificial intelligence
- 8:35eventually replace the philosopher?
- 8:38I'll generate it. This is what it looks
- 8:40like.
- 8:42Can artificial intelligence eventually
- 8:44replace the philosopher?
- 8:46So now, if I generate it a second time,
- 8:49this is what it looks like.
- 8:51Can artificial intelligence eventually
- 8:54replace the philosopher?
- 8:58One last time, can artificial
- 8:59intelligence eventually replace the
- 9:01philosopher?
- 9:04So you see, it's really very random,
- 9:06which means that it takes a lot of time
- 9:09to regenerate the sentences and get
- 9:11sentences that sound natural.
- 9:14And so, it's obvious that what I'm
- 9:16saying here is valid for today, that is
- 9:19to say in 2025. In other words, if you
- 9:21listen to this episode in 2030 or 2040,
- 9:25you'll probably be thinking, "What is
- 9:27this AI that doesn't place the
- 9:28intonations where they should be?"
- 9:30Because it is very likely that in a few
- 9:33years, LIA will be able to correctly
- 9:36place voice intonations. "And so at that
- 9:39point, we will look at the AI of 2025 as
- 9:42the prehistory of AI.
- 9:44So that was for the voice.
- 9:47Now, regarding the script.
- 9:50So, for the script, we generated it with
- 9:52chat GPT version 4. O, I say we because
- 9:55in reality, it was mainly my partner
- 9:57David who did this work of generating
- 9:59the script.
- 10:02And so obviously the goal of the game
- 10:04was not to modify anything manually, it
- 10:06was not to add anything, not to reword
- 10:08anything even slightly because otherwise
- 10:10, well, it would be cheating, it would
- 10:12be too easy.
- 10:14So the idea was that the entire script
- 10:17would be generated by AI.
- 10:19So, we did allow ourselves two things.
- 10:22We allowed ourselves to remove sentences
- 10:24because we didn't want it to be too long
- 10:25.
- 10:27And we also allowed ourselves to
- 10:29rearrange the script, that is to say to
- 10:32change the order of certain sentences to
- 10:34make the message more fluid.
- 10:37But I really insist on this on that
- 10:38without changing anything in the text.
- 10:41And so, as you know, when you want Chat
- 10:44GPT to generate a text, you have to give
- 10:46it instructions. You have to write what
- 10:49is called a prompt.
- 10:51And so the prompt, from what I know,
- 10:53from what I have understood, is really
- 10:55the key.
- 10:56It is really what will determine the
- 10:59quality of the result.
- 11:01So here is the prompt that was used to
- 11:03generate the script. I am displaying it
- 11:05on the screen for those who are
- 11:06listening on YouTube.
- 11:08I would like you to write the script for
- 11:10an audio podcast in the style of Charles
- 11:13Robin, aka the tutor.
- 11:15The topic will be: Can artificial
- 11:17intelligence replace the philosopher?
- 11:21The script should be around 1,500 words
- 11:23and you should stick as close as
- 11:25possible to Charles 'style.
- 11:27The same tone, the same turns of phrase
- 11:30, the same argumentative approach and
- 11:32the same mannerisms.
- 11:34The goal is for the audience to be aware
- 11:36, so that they do not realize that the
- 11:38script was generated by UNIA, but that
- 11:40they think it is Charles. himself who
- 11:43wrote it. I can provide you with several
- 11:45scripts of these episodes so that you
- 11:48can analyze his writing style and get as
- 11:50close as possible to it.
- 11:53Now, I want to make it clear that I have
- 11:54absolutely no idea whether this prompt
- 11:56is good or not.
- 11:57What I mean by that is that writing
- 12:00prompts is a skill.
- 12:02There is a way of writing the prompt
- 12:04that determines whether or not you will
- 12:06get a good result.
- 12:08However, it turns out that neither my
- 12:10partner David nor I have this skill.
- 12:14So, we did what seemed most relevant to
- 12:16us.
- 12:17But if you have any comments or
- 12:19suggestions for improving the prompt, to
- 12:21get a better result, feel free to tell
- 12:23me in the comments.
- 12:25I will read it with great interest.
- 12:27And so, here we are with a first version
- 12:30of the script.
- 12:32So, the first observation may seem
- 12:34anecdotal, but personally, I find it
- 12:36very revealing.
- 12:38I ask Chat GPT to generate a script that
- 12:40perfectly imitates my style.
- 12:43I give it scripts of my episodes so that
- 12:46it can identify my style.
- 12:49And in the script that she generates for
- 12:51me, she doesn't include the" hello
- 12:53everyone "which is kind of the trademark
- 12:55of all my episodes.
- 12:58So the" hello everyone "that you heard,
- 13:00I'm telling you because I want to be
- 13:01completely transparent, I added it
- 13:02manually.
- 13:05I took it from another episode and I put
- 13:07it at the beginning of what you listened
- 13:09to because the goal was to make you
- 13:11believe that it was me.
- 13:14So if you see something wrong from the
- 13:16first sentence, well it couldn't work.
- 13:19Same thing for the" thank you "at the
- 13:21end.
- 13:22Chat GPT didn't include it. So there too
- 13:25, I added it manually but except that
- 13:27unlike the" hello everyone ", I didn't
- 13:29take it from another episode, I
- 13:31generated it on Eleven Labs.
- 13:34So, I don't know what you think, but
- 13:37personally, the fact that this" hello
- 13:40everyone "and this" thank you "which,
- 13:42once again, are really a constant in all
- 13:45In my episodes, the fact that chat GPT
- 13:48did not identify these two recurring
- 13:51phrases as an integral part of my style
- 13:54, I find that very surprising, as if
- 13:56what was obvious and obvious to us
- 13:59humans was not at all so to her.
- 14:02And then even worse, if I dare say so,
- 14:05in the script generated by chat GPT, we
- 14:08spotted an error, a factual error that I
- 14:11deliberately left in and that some of
- 14:14you may have noticed. I'm talking about
- 14:17the quote attributed to Simon Veille,
- 14:20thinking is saying no.
- 14:22In reality, this quote is not from Simon
- 14:24Veille, it is from the philosopher Alain
- 14:26.
- 14:27And why do I find that crazy?
- 14:29Because on the one hand, chat GPT
- 14:31manages to do incredible things. It
- 14:34manages to generate coherent texts, to
- 14:37generate reasoning, to imitate a writing
- 14:40style.
- 14:41But at the same time that it manages to
- 14:43do all that, it makes gross errors,
- 14:46errors that could be avoided with two
- 14:48clicks on Google.
- 14:50Even though chat GPT has access to all
- 14:52the resources available on the internet
- 14:54.
- 14:56So if you have an explanation Tell me
- 14:58about this incongruity because I have to
- 15:00admit that I'm stuck here.
- 15:04So if we put that aside in terms of
- 15:06writing style, I found that it was still
- 15:08quite similar, that it was quite close
- 15:10to my style.
- 15:12Well, there were still moments when I
- 15:13said to myself," You can clearly see
- 15:15that it's not me. "I would never have
- 15:16said it like that.
- 15:18For example, the sentence with a
- 15:20fluidity and clarity that can make you
- 15:22dizzy.
- 15:24That's not typically my way of writing.
- 15:26Making you dizzy is the kind of
- 15:28expression I never use. There are also
- 15:31passages that I found a bit disjointed,
- 15:33ideas that followed on from one another
- 15:35in a somewhat clumsy way.
- 15:37But despite its flaws, I would say that
- 15:39if you're not aware of it, you can still
- 15:42fall for it.
- 15:43So that's the technical aspect. Now I
- 15:46come to the strictly philosophical
- 15:49aspect.
- 15:50So on a philosophical level, I have
- 15:51quite a few things to say. say.
- 15:54The first question that needs to be
- 15:56asked is, did Lia do the job?
- 15:59Did Lia succeed in generating
- 16:01philosophical reflection?
- 16:04Personally, I would tend to answer yes.
- 16:08I would tend to answer yes in the sense
- 16:10that she did indeed develop a
- 16:12philosophical argument which is also the
- 16:15classic argument when we want to support
- 16:18the idea that Lia does not think, namely
- 16:21the argument of affect, the argument of
- 16:23feeling.
- 16:25She does not feel and because she does
- 16:28not feel, she does not think in the
- 16:31sense that thought is not a process
- 16:33independent of affect, but on the
- 16:36contrary an emergence of affect.
- 16:39This is an argument that I find
- 16:41interesting.
- 16:42I am not saying that it is valid, I am
- 16:44saying that it deserves to be considered
- 16:46.
- 16:47Especially if you are a materialist,
- 16:49that is to say, someone who considers
- 16:51that thought always has a material
- 16:53substrate.
- 16:55Ideas are not floating entities, they
- 16:58are productions of the body with all
- 17:01that that implies in terms of the life
- 17:04of the body.
- 17:06And so obviously, I think you see the
- 17:08paradox.
- 17:10The paradox is the fact that Lia is able
- 17:13to say of herself that she does not
- 17:15think, that she cannot think because she
- 17:17does not feel.
- 17:21In other words, Lia is able to know what
- 17:23her very structure is supposed to
- 17:25prevent her from knowing, namely that
- 17:27she is not capable of thinking.
- 17:31And that is why I think that once we
- 17:33have said that, once we have said that
- 17:35AI does not really think, we must move
- 17:38on to the next step.
- 17:40We must move on to the next question
- 17:43which is not can AI think but can AI
- 17:46replace the thinker?
- 17:49Can it replace the philosopher?
- 17:51So where is the difference?
- 17:55Well the difference, pay attention, is
- 17:58that we only know of a thought what is
- 18:00expressed by it.
- 18:04A human being who thinks, we never
- 18:06really have access to his thought.
- 18:10We only have access than to what he says
- 18:12about his thoughts.
- 18:15We have access to his speech, to his
- 18:17writings, that is to say, to his words.
- 18:21And it is because we know, I should say
- 18:24, it is because we assume that words are
- 18:26always the expression of something, that
- 18:29words are always the expression of a
- 18:32thought in the background of the words
- 18:35that we deduce that the human being who
- 18:38expresses himself thinks.
- 18:42But how do we know that the human being
- 18:44who expresses a thought really thinks?
- 18:49How do we know that there is a mental
- 18:52activity, an inner activity, therefore
- 18:55invisible, upstream of the expression of
- 18:58this activity?
- 19:01Here, for example, I am talking to you.
- 19:04I am expressing my thoughts.
- 19:08But the only thing you have access to
- 19:11are my words.
- 19:14Not the thoughts that are manifested in
- 19:16my words.
- 19:17You assume that I have a thought because
- 19:20you assume that I am a human being.
- 19:24But in reality, what do you know about
- 19:27it?
- 19:28What tells you that I am not a robot?
- 19:32What tells you that I really think?
- 19:36Of course I really think, but you only
- 19:39know that from what I tell you.
- 19:42We judge the existence of a thought by
- 19:44the expression of that thought.
- 19:47In other words, we judge a thought by
- 19:49its effects.
- 19:51We judge a thought by its appearance.
- 19:54Words are the appearance of thought.
- 19:58And so if we stick only to the
- 20:01appearances of thought, if we stick only
- 20:04to words and writings, we cannot
- 20:07prejudge anything about thought as an
- 20:10internal process behind expression.
- 20:14Can someone think?
- 20:16Personally, I would say no.
- 20:19But that is not the question.
- 20:22The question is, can AI replace the
- 20:25philosopher?
- 20:27And replacing the philosopher means
- 20:30producing the effects of thought.
- 20:33If AI can produce the effects of thought
- 20:36, then it can in fact replace the
- 20:38philosopher.
- 20:41The question Lia thinks is not a good
- 20:44question in reality because This is an
- 20:46unverifiable question.
- 20:49We never judge it,
- 20:52we always judge the result of thought.
- 20:57In other words, the real question is not
- 20:59, does Lia think?
- 21:02The real question is, can Lia imitate to
- 21:04the point where we can no longer detect
- 21:07it?
- 21:09The real question is that of detection.
- 21:13It is that of the possibility of
- 21:16identifying a speech, of identifying a
- 21:19text as having been generated or not by
- 21:21a human being because that is the only
- 21:24thing we have, because it is the only
- 21:27question we can answer.
- 21:30What if Lia is capable of generating
- 21:32texts that we think were generated by a
- 21:35human being or that we are unable to
- 21:37know whether or not they were generated
- 21:40by a human being?
- 21:42To the question," Can Lia replace the
- 21:45philosopher? "The answer is yes.
- 21:48Let's do a thought experiment.
- 21:52Imagine that a philosophy book is
- 21:54published anonymously.
- 21:57We do not know its author.
- 21:59But this book provokes unprecedented
- 22:02enthusiasm.
- 22:03It completely revolutionizes our usual
- 22:06way of conceiving the world.
- 22:09After reading it, you might say to
- 22:11yourself:" This book is incredible!
- 22:14I've never read anything so powerful, so
- 22:18moving, so convincing. "Now imagine that
- 22:22we end up learning that this book was
- 22:25written by Unia.
- 22:27What would you say?
- 22:30Would you say," Well, maybe this book
- 22:33isn't that extraordinary after all.
- 22:37Maybe I got a little carried away. "" If
- 22:40you tell yourself that, you are making
- 22:43what is called a retrospective judgment
- 22:45.
- 22:47You give credit to this book and when
- 22:50you learn that it was not written by a
- 22:52human, you withdraw that credit.
- 22:55There is a name for this.
- 22:58Cheating.
- 22:59You are cheating because you are
- 23:01changing the rules during the game.
- 23:06Either the book is good, relevant,
- 23:08convincing, and in this case it remains
- 23:10so even if you learn that it was written
- 23:13by a human, or it is not.
- 23:17And if it is not, you have to realize it
- 23:20before you know that it was written by a
- 23:22human.
- 23:24I insist, the criterion of detection.
- 23:27And so that is where we come to the
- 23:29question of imitation.
- 23:32Lia does not think, she limits thought.
- 23:36That is true.
- 23:38But what difference does it make?
- 23:41What difference does it make to you?
- 23:43Whether the text you have before you is
- 23:46the product of a mental activity or the
- 23:48imitation of a mental activity?
- 23:52What difference does it make to you?
- 23:52Does that change anything? Because in
- 23:54both cases, you only have access to the
- 23:57result of this activity, never to the
- 23:59activity itself.
- 24:02If you go to a museum, you come across a
- 24:05painting, and it moves you to the point
- 24:08of shedding a tear, and then you learn
- 24:11that this painting was generated by an
- 24:14AIAI, how does that change the aesthetic
- 24:18experience you had?
- 24:21Absolutely nothing.
- 24:23An emotion is never false.
- 24:27An emotion provoked by an imitation is
- 24:30nonetheless a real emotion.
- 24:33Well, in the same way, Lia imitates the
- 24:36thought.
- 24:37But if her imitation produces the same
- 24:40effects as an original thought, if her
- 24:43imitation is so convincing that it
- 24:45becomes indistinguishable from the
- 24:48original, she fulfills her mission,
- 24:50because we judge a thought not by the
- 24:53inner life of the person who produced it
- 24:56, but by the effects it provokes on the
- 24:59person who receives it.
- 25:03In an interview he gave on the Studio
- 25:05Beggle channel, Alexandre Astier spoke
- 25:08about literary creation and he said that
- 25:10in Motant he was talking about the
- 25:13inhabitants of Rouan and that he called
- 25:15them the pointed people of the Gothic
- 25:18bell towers.
- 25:20If it was an AI that wrote the pointed
- 25:23people of the Gothic bell towers, I find
- 25:26it much less interesting.
- 25:29Well, in this case, it's this sentence
- 25:32that is interesting because Alexandre
- 25:34Astier doesn't say "If an AI wrote that
- 25:36, I find it less good."
- 25:40And fortunately, by the way, he doesn't
- 25:42say that, because if he did say that, he
- 25:44would find himself in the contradiction
- 25:46that I was just talking about: namely,
- 25:48giving value to a text only to
- 25:50ultimately remove that value when we
- 25:52learn that it wasn't written by a human
- 25:54.
- 25:56He says "I find that less interesting."
- 26:00less interesting because it means that
- 26:04behind the text, there is not what we
- 26:08like to see behind a text, namely an
- 26:11affect, an emotion, an experience.
- 26:15There is no inner being that this
- 26:19sentence is supposed to reflect.
- 26:23Everything is there.
- 26:26Someone who would say: "If this sentence
- 26:28had been written by someone, I would
- 26:31find it less good, less beautiful, less
- 26:33poetic" is someone who would be saying:
- 26:36"I am not happy, I am not happy to have
- 26:39been had and I am expressing my
- 26:41discontent by a retrospective judgment
- 26:43that says the opposite of what I think
- 26:46because I do not want to admit that what
- 26:48Unia writes could touch me."
- 26:53Human beings love mystery.
- 26:58They love the idea that there is
- 27:00something hidden, invisible,
- 27:02inaccessible.
- 27:05So much so that they place more value on
- 27:08what seems to them to escape all
- 27:10explanation than on what works
- 27:12transparently.
- 27:15Human beings find charm to that which
- 27:19escapes causality.
- 27:22Now, a machine is a system of cause and
- 27:26effect of open and closed circuits of
- 27:30zero and 1. Whereas humans are
- 27:34unpredictable, enigmatic, and
- 27:37inexplicable.
- 27:40If we tend to attach less value to a
- 27:43thought generated by I than to a thought
- 27:46generated by a human, it is because deep
- 27:50down inside us there is a love of the
- 27:53unknown, a love of the inexpressible, a
- 27:56love of that "je ne sais quoi" which,
- 27:59because it eludes all explanation,
- 28:01satisfies our need for things to be
- 28:04inexplicable.
- 28:07Because if everything is explainable,
- 28:09what is left for us to conquer?
- 28:13What am I getting at?
- 28:15I want to get to this question.
- 28:18What tells us that the je ne sais quoi
- 28:22of the human being is not simply a
- 28:26matter of complexity and not No mystery
- 28:30?
- 28:33To speak of mystery presupposes a
- 28:36qualitative difference between man and
- 28:39machine, a difference in nature.
- 28:45To speak of complexity presupposes a
- 28:48quantitative difference, a difference in
- 28:51degrees.
- 28:54Perhaps we are simply extremely complex
- 28:57machines.
- 29:00So complex that we have ended up
- 29:03attributing to ourselves a mysterious,
- 29:06impenetrable, almost divine essence.
- 29:12Perhaps what we call mystery is nothing
- 29:14more than complexity that has yet to be
- 29:16resolved.
- 29:19We call mystery what we do not know.
- 29:23We call human what seems beyond the
- 29:26reach of the machine.
- 29:30What will happen the day the machine
- 29:33takes this step?
- 29:37The day when, without us really knowing
- 29:40how, its quantitative progress takes the
- 29:43form of a qualitative change.
- 29:48The day when the neural network that
- 29:52composes it becomes so rich, so
- 29:55extensive, so complex, that the human
- 30:00brain will seem rudimentary in
- 30:03comparison.
- 30:06The day when we understand that the real
- 30:08question is not can the machine replace
- 30:10the human?
- 30:14Rather, how much time is left to the
- 30:16human machine before being overtaken by
- 30:20a more complex machine?
- 30:24I know what people will say to me.
- 30:26That it won't happen tomorrow, that the
- 30:30machine will never be able to even
- 30:33approach what constitutes the essence of
- 30:37human uniqueness, to vibrate as we can
- 30:41vibrate, to love as we can love, to
- 30:44desire as we can desire.
- 30:48Perhaps, but perhaps also that we don't
- 30:52know anything about it, that all this is
- 30:55for the moment, pure speculation, that
- 30:58pure speculation because, as so often in
- 31:02history, we will understand things when
- 31:05it is too late to change them.
- 31:09In the meantime, beware of our pride,
- 31:13this pride that for the moment remains
- 31:17the privilege of human beings and that
- 31:21makes us blind not only to what is
- 31:24hidden but to what we do not want to see
- 31:28.
- 31:30Thank you, and this time it was me or
- 31:33not.
About this transcript
This page contains the full transcript of INTELLIGENCE ARTIFICIELLE - Peut-elle remplacer le philosophe ? by Le Précepteur, generated from the public captions YouTube serves with the video. The transcript has 4,303 words across 715 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.